U.S. patent application number 17/838549 was filed with the patent office on 2022-09-29 for two-wheel in-line robots.
The applicant listed for this patent is TWILL Technology, Inc.. Invention is credited to Cameron Tacklind, Christopher Andrew Tacklind.
Application Number | 20220308585 17/838549 |
Document ID | / |
Family ID | 1000006390947 |
Filed Date | 2022-09-29 |
United States Patent
Application |
20220308585 |
Kind Code |
A1 |
Tacklind; Christopher Andrew ;
et al. |
September 29, 2022 |
Two-wheel In-Line Robots
Abstract
Techniques are disclosed for exploiting modern controls, sensors
and actuators to realize a novel family of in-line two-wheeled
vehicles (Twills) as robots. Each robot has a front-wheel with a
substantially horizontal axis of rotation and a substantially
vertical steering axis. The front-wheel with its substantially
vertical steering axis has a steering-angle that can be sensed by
one or more sensors. There is a rear-wheel with a substantially
horizontal axis of rotation. A control module stabilizes the roll
angle when the robot is in a forward motion as well as when it is
substantially or fully stopped. One or both the wheels of the robot
may be endowed by a steering motor for steering and a traction
motor for providing traction/torque to the wheel.
Inventors: |
Tacklind; Christopher Andrew;
(Menlo Park, CA) ; Tacklind; Cameron; (Menlo Park,
CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
TWILL Technology, Inc. |
Menlo Park |
CA |
US |
|
|
Family ID: |
1000006390947 |
Appl. No.: |
17/838549 |
Filed: |
June 13, 2022 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
16901333 |
Jun 15, 2020 |
11392126 |
|
|
17838549 |
|
|
|
|
15214258 |
Jul 19, 2016 |
10739772 |
|
|
16901333 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
B62K 11/007 20161101;
G05D 2201/0212 20130101; G05D 1/021 20130101 |
International
Class: |
G05D 1/02 20060101
G05D001/02; B62K 11/00 20060101 B62K011/00 |
Claims
1. A robot comprising a body to which are mounted: (a) a
front-wheel and a rear-wheel; (b) a first steering motor for
steering a first wheel from amongst said front-wheel and said
rear-wheel, wherein said first wheel has a substantially vertical
steering axis; (c) a first traction motor for a second wheel from
amongst said front-wheel and said rear-wheel; (d) a controller
which stabilizes a roll angle while said robot is in forward motion
by adjusting a steering-angle in response to said roll angle; and
(e) a controller which stabilizes said roll angle while said robot
is substantially stopped by setting said steering-angle to a
substantially nonzero position and by applying a torque to said
second wheel about its axis of rotation.
2. The robot of claim 1 further comprising a second steering motor
for said second wheel.
3. The robot of claim 1 further comprising a second traction motor
for said first wheel.
4. The robot of claim 1 further comprising an encoder for encoding
said steering-angle.
5. The robot of claim 1 wherein said encoder is based on a
technology including magnetic and optical technologies.
6. The robot of claim 1 further comprising a suspension system that
allows a vertical motion of said body while said robot is in
forward motion, and relax-dropping said body to the ground when
said robot is stopped.
7. The robot of claim 6 wherein said suspension system comprises a
pressurized hydraulic source which when pressurized enables said
vertical motion and when relieved enables said relax-dropping.
8. A method comprising the steps of: (a) attaching a front-wheel
and a rear-wheel to a body of a robot; (b) providing a first
steering motor for steering a first wheel from amongst said
front-wheel and said rear-wheel, with said first wheel having a
substantially vertical steering axis; (c) providing a first
traction motor for a second wheel from amongst said front-wheel and
said rear-wheel; (d) providing a controller for stabilizing a roll
angle while said robot is in forward motion by adjusting a
steering-angle in response to said roll angle; and (e) providing a
controller for stabilizing said roll angle while said robot is
substantially stopped by setting said steering-angle to a
substantially nonzero position and by applying a torque to said
second wheel about its axis of rotation.
9. The method of claim 8 further providing a second steering motor
for said second wheel.
10. The method of claim 8 further providing a second traction motor
for said first wheel.
11. The method of claim 8 further providing an encoder for encoding
said steering-angle.
12. The method of claim 11 selecting said encoder from a technology
including magnetic and optical technologies.
13. The method of claim 8 further providing a suspension system for
enabling a vertical motion of said body while said robot is in
forward motion, and for relax-dropping said body to the ground when
said robot is stopped.
14. The method of claim 13 further providing said suspension system
to include at least one pressurized hydraulic source which when
pressurized enables said vertical motion and when relieved enables
said relax-dropping.
Description
RELATED APPLICATIONS
[0001] This application is a continuation of and claims priority
from now allowed U.S. patent application Ser. No. 16/901,333 filed
on Jun. 15, 2020, which is a continuation of U.S. Pat. No.
10,739,772 B2 issued on Aug. 11, 2020. The above numbered patent
application and patent are incorporated by reference herein for all
purposes in their entireties.
FIELD OF THE INVENTION
[0002] This invention generally relates to in-line two-wheeled
vehicles (Twills) operating as robots.
BACKGROUND ART
[0003] Fully autonomous vehicles are clearly on the horizon. Major
companies are rushing forward with self-driving cars. The CEO of
Uber has made it clear that they eventually intend to have fully
automatic vehicles that deliver packages and humans. Is a vehicle
that delivers a human a "car" or a robot? Still others consider
using robotic contrivances to deliver packages. Notably, Amazon and
others propose using robotic drones for delivery. Others are
demonstrating small robots for to-the-door delivery.
DISCUSSION OF PRIOR ART
[0004] For many if not most applications, ground delivery is more
efficient than air-borne delivery systems. In the drive to make
package delivery cheaper and more effective, ground delivery will
dominate. Hence ground robots will be pressed into service as
delivery robots.
[0005] There is a plethora of robot configurations. The
"Differential drive" is common for indoor robots on flat floors.
They have difficulty on thresholds and any other floor
imperfection. "Ackerman steering" is the arrangement common in
cars. This works well in robots and is more forgiving on irregular
surfaces. "Treaded" robots are often used in more severe
environments. They are limited in their top speed and
efficiency.
[0006] Like most robots, these are all limited in height. The
primary characteristic dimension of a robot is the shorter of the
wheel base length and wheel base width. Generally, the height is
limited by this characteristic dimension. In more plain terms, tall
robots need to be wide so that they don't fall over. The more
extreme the terrain, the more the height is limited.
[0007] Balancing robots such as the now common "Segway.TM. robots"
get around this limitation by virtue of their "inverted pendulum"
control dynamics. In particular, if a balancing robot is made 4
times taller, the rotational dynamics are half as fast, making the
balancing servo control required half as fast. While clever, these
all have severe limitations in control authority while moving. In
particular it is noted that servo motors have maximum torque at
zero speed. So, they have impressive control authority at zero
speed. But as the speed increases, the torque available drops. So
as a balancing robot speeds up, it loses control authority. In
particular, the only way for a balancing robot to slow down is to
get the wheels out in front of the center of mass. So, if the
maximum speed of the motors is reached, there is no control
authority. In plain terms, it falls down.
[0008] Segway.TM. robots have an additional limitation in "roll".
In particular, the characteristic dimension of the wheel spacing
limits the height if the ground surface is not level side-to-side.
Some versions add an additional degree of freedom to correct for
roll displacement of the center of mass, but this adds complexity
and puts additional limits on the roll stability.
[0009] The prior art does include two in-line wheeled vehicles that
are stabilized by using torque from a large mechanical gyroscope.
However, these are heavy archaic vehicles and are collectively
known as Gyrocars. Ford Motor Company produced sample Gyrocars in
the 1960s. Other art describes a wide variety of automatic kick
stands and training wheels to catch the vehicle as it stops.
OBJECTS OF THE INVENTION
[0010] In view of the shortcomings of the prior art, it is an
object of the invention to provide modern controls, sensors and
actuators to realize a novel family of in-line two-wheeled vehicles
(Twills) as robots.
[0011] Other objects and advantages of the invention will become
apparent upon reading the summary and the detailed description in
conjunction with the drawing Figures.
SUMMARY OF THE INVENTION
[0012] The invention at hand exploits modern controls, sensors and
actuators to realize a novel family of in-line two-wheeled vehicles
(Twills) as robots. Each robot has a front-wheel with a
substantially horizontal axis of rotation and a substantially
vertical steering axis. The front-wheel with its substantially
vertical steering axis has a steering-angle that can be sensed by
one or more sensors. There is a rear-wheel with a substantially
horizontal axis of rotation. A control module stabilizes the roll
angle when the robot is in a forward motion or when it is stopped.
There is an energy source to power the controls and to drive one or
both the front and rear wheels.
[0013] These two-wheel in-line vehicles are ideally suited for a
wide variety of applications beyond humans driving a Twill. As
these vehicles become more ubiquitous and automatic, they will
inevitably be used to deliver/transport humans.
[0014] In the apparatus and methods of various embodiments, there
is a steering motor for steering one of the front-wheel and the
rear-wheel, also referred to as the first wheel. There is also a
traction motor for driving the other wheel from amongst the front
and the rear wheels, also referred to as the second wheel. There is
a controller or control module that stabilizes the roll angle of
the robot when the robot is in forward motion.
[0015] The controller does so by adjusting the above-discussed
steering-angle in response to the roll angle. The controller
further stabilizes the roll angle when the robot is substantially
stopped. It does so by setting the steering-angle to a
substantially nonzero position and by applying a torque via the
above traction motor to the second wheel.
[0016] In a variation, there is also a second steering motor
provided for the second wheel. In another or the same variation, a
second traction motor is also provided for the first wheel. In
variations, the steering-angle is encoded via an encoder. The
encoder may comprise technologies of the art including magnetic and
optical technologies.
[0017] In systems and methods of still other embodiments, there is
a suspension system that moves the body of the robot up and down
i.e. in a vertical motion, while the robot is in forward motion.
The same suspension system relax-drops the robot body to the ground
when the robot is stopped.
[0018] In the same or related embodiments, the suspension system
comprises a pressurized hydraulic source. Pressurizing the
hydraulic source enables the above-mentioned vertical motion of the
robot body. Relieving or depressurizing the hydraulic source
enables the above-mentioned relax-dropping of the body of the robot
to the ground.
[0019] Clearly, the system and methods of the invention find many
advantageous embodiments. The details of the invention, including
its preferred embodiments, are presented in the below detailed
description with reference to the appended drawing Figures.
BRIEF DESCRIPTION OF THE DRAWING FIGURES
[0020] FIG. 1 shows an oblique view of a Twill bot demonstrator for
delivering manila folders.
[0021] FIG. 2 shows an oblique view of a compact steering mechanism
of the prior art.
[0022] FIG. 3 shows a top view of the prior art steering mechanism
indicating the instant center concept.
[0023] FIG. 4 shows a top view of the prior art with a non-zero
steering-angle.
[0024] FIG. 5 shows a side view of a novel extension of this
steering mechanism with hydraulic synchronization.
[0025] FIG. 6 shows the interconnections between hydraulic
cylinders to produce synchronous motion and damping.
[0026] FIG. 7 shows and improved embodiment that substantially
eliminates bump steering.
[0027] FIG. 8 shows a Twill delivery robot suitable for on-road
use.
[0028] FIG. 9 shows a Twill delivery robot with an integral arm
mechanism for loading and unloading.
[0029] FIG. 10 shows an upward view of a low-profile mule
robot.
[0030] FIG. 11 shows an oblique view of a mule robot deploying a
tall sensor array.
[0031] FIG. 12 shows a Twill robot configured as a telescoping
telepresence robot.
DETAILED DESCRIPTION
[0032] The figures and the following description relate to
preferred embodiments of the present invention by way of
illustration only. It should be noted that from the following
discussion, alternative embodiments of the structures and methods
disclosed herein will be readily recognized as viable alternatives
that may be employed without departing from the principles of the
claimed invention.
[0033] Reference will now be made in detail to several embodiments
of the present invention(s), examples of which are illustrated in
the accompanying figures. It is noted that wherever practicable,
similar or like reference numbers may be used in the figures and
may indicate similar or like functionality. The figures depict
embodiments of the present invention for purposes of illustration
only. One skilled in the art will readily recognize from the
following description that alternative embodiments of the
structures and methods illustrated herein may be employed without
departing from the principles of the invention described
herein.
[0034] FileBot
[0035] We assert that twill robots are suitable for a wide variety
of tasks. A major category is the task of ground delivery. As an
example of this utility we have reduced to practice a simple
delivery robot that delivers manila folders (15). To accentuate the
novel features of this robot, we fabricated this demonstration as a
double size manila folder (13). FIG. 1 shows such a "fileBot"
standing with its wheels (1F, 1R) turned sideways. A volume (16) is
provided for electronics, sensors, motors and batteries.
[0036] In this preferred embodiment, the fileBot has two identical
legs (11) disposed as far apart from each other as is practical.
Each leg is fitted with a turning servo (12F, 12R) that acts to
twist the leg relative to the body (13) and a torque motor (10)
that drives the wheel at the bottom of the leg. In another
preferred embodiment, only one leg may twist and only one-wheel has
a torque motor. In the preferred embodiment shown, the contact
point of the wheel is on axis with the axis of rotation of the leg.
An offset of this contact point is known as trail. For the purposes
of this first demonstrator, the trail adds unneeded complexity to
the equations. Likewise, making the steering-angle non-zero is
common, but not needed for this simple demonstration.
[0037] Segway.TM. Style Balance
[0038] In this stance the fileBot balances just as a Segway.TM.
robot does. Though limited in speed, it has considerable control
authority at low speeds. Balancing a fileBot near zero speed is now
commonly practiced by persons skilled in the art. It has literally
the dynamics of an inverted pendulum. The control law is
fundamentally a "PID controller" around the pitch axis. That is,
the wheels move forward and aft relative to the center of mass in
an attempt to keep the contact point on the ground substantially
under the center of mass. The estimate for pitch and first
derivative of pitch are readily supplied by modern accelerometers
and rate gyros. "Sensor fusion" is typically achieved by use of a
Kalman filter or simpler digital filter.
[0039] This fileBot demonstrator incorporates half a dozen optical
range finders (14F, 14R). So, it is possible for the fileBot to
align itself parallel to a wall. This is particularly handy for
parking. A fileBot can readily roll up to a wall and with a tiny
flick of the wheels, fall over a small amount to come to rest
against the wall. The reverse is also true. By rolling into the
wall, enough angular momentum is generated to pull the fileBot away
from the wall. At the next moment, the balancing algorithm is
activated arresting the pitch and holding the pitch angle at zero.
A wide variety of other sensors are also applicable.
[0040] Segways Crash from being Wide
[0041] A fileBot moving in this Segway.TM. manner will be
impressive enough. But as it moves in this manner it has a very
wide stance. This is a rather poor choice for moving through a
cluttered environment. Like any Segway.TM. robot, it would suffer
considerably as the speed increases. If one edge clips a stationary
object it is certain to crash to the floor.
[0042] Pitch Definition
[0043] Note that in the discussion above, pitch was defined
relative to the "Segway.TM. like" motion with the wheels turned
transverse to the body of the fileBot. The motion of a fileBot is
much more interesting along the axis of the long dimension.
Defining this as the X direction it is convenient to define X, Y,
and Z, and pitch, roll, and yaw in the conventional sense. The
challenge in balancing a fileBot while moving in the forward X
direction is to stabilize the roll angle. Note that for the most
part, no stabilization is needed in pitch as the characteristic
dimension (17) of the spacing of the wheels provides pitch
stiffness along the X direction.
[0044] Forward Motion
[0045] In this forward motion case, the dynamics are still that of
an inverted pendulum. That is, the challenge is to stabilize the
roll by moving the contact point in the Y direction relative to the
center of mass. This must happen while the X position is steadily
increasing.
[0046] Note that using the torque motors as in the Segway sense,
only puts ripple in the X velocity. It has little to no effect on
the roll angle. Instead, we effect roll by commanding a non-zero
steering-angle (16). Since the forward momentum is used to affect
this roll behavior, the control authority increases with forward
speed.
[0047] Two-Wheel Steering
[0048] For these experiments it is worth considering steering the
wheels together. This keeps the X direction substantially constant
instead of introducing yaw motion. For a particular velocity, a
small steering-angle will cause the contact patch of the wheels to
creep out from under the fileBot. The Roll rate is then directly
proportional to the forward velocity and the steering-angle.
[0049] Slope of Steering
[0050] More precisely, it is the "slope" of the steering-angle that
is the relevant number. For a given forward velocity, doubling the
steering-angle slope will double the roll rate. For the PID
controller one skilled in the art will readily find suitable gain
constants commanding the slope in proportion to the roll and roll
rate ("PD controller"). Further note that if the forward speed is
cut in half, the steering slope needs to double to create the
equivalent control influence. By extrapolation we get the natural
result that as the speed goes to zero, the steering slope goes to
infinity. This is exactly the Segway.TM. condition. This example
provides a first preferred embodiment of the control law that spans
from forward motion to a fixed stance.
[0051] Other Control Methods--Segway.TM. Mode
[0052] As the development of Twills advances, many variants of the
control laws will be employed. Take for example the control law for
standing still in the Segway.TM. mode. In this stance, the steering
motors may hold the wheels perpendicular to the body. Then the
challenge is to command torque to the traction motors in response
to the roll angle and roll rate. In a preferred embodiment, the
side-to side displacement distance Y and rate of change of Y are
also taken into consideration. One is tempted to use a control law
where the wheel torque is proportional to these four variables.
Numerous simulation examples are available of this technique. Many
sophisticated analysis techniques are also available. Even methods
for algorithmically generating the proportionality constants are
readily available. In practice however, one needs to deal with
practical limits such as back lash, noise, and saturation of the
actuators. Failure to account for these leads to erratic behavior
and quickly to complete failure of the balancing function.
[0053] Multiple Loops
[0054] So, it is often preferable to break the problem into simpler
independent loops. In a preferred embodiment, a fast servo loop may
be formed controlling the position of the wheels in the Y
direction. This is readily formed by a PID loop on the sensed
position of the wheels. This control may have any selected dynamic
behavior from snappy to sluggish. It is common practice by those
skilled in the art to optimize for minimum settling time. In as
much as this commanded position is fast compared to the roll
dynamics, it can be considered to be instantly available. So, in
this preferred embodiment, the Y position may be commanded in
direct proportion to the roll angle and roll angle rate. Adding an
integral term results in the common PID construct. Now suppose we
desire the Y position to be zero. All we need to do is form another
PID loop on Y which commands the roll angle. This is how it comes
about that "to move home you need to lean to home. To lean to home,
you must drive away from home." Still other loops may be added to
account for sensor drift.
[0055] Forward Motion Servo
[0056] Note that in the "slope of steering" teaching above, a servo
loop was used to control the steering-angle while driving. In a
preferred embodiment, a control loop on side-to-side motion may be
formed again, controlling the Y position. That is, if a 0.1 m step
is commanded, the slope is commanded to increase and then decrease
as the correct position is achieved. Note that the computation of
the Y position also takes into account the forward velocity. That
is, if you are going twice as fast, you get to the required Y
displacement twice as fast. As in the Segway.TM. mode, we can now
consider the Y value as fast compared to the roll error. So we
again command the Y position in response to the Roll error forming
a PID controller for the roll. In a preferred embodiment,
additional loops may be imposed as needed. Note that in this second
preferred embodiment the control law for forward motion is
essentially identical to the control law for Segway.TM. mode. In
preferred embodiments the transition from one mode to the other may
be given special consideration.
[0057] Crucial Benefits
[0058] This use of the forward momentum is a crucial benefit of the
invention. It is this feature of the invention that results in high
control authority at working velocities. The invention further
requires the transition to balancing while stopped. The invention
also requires the transition from balancing to forward motion. This
is also the crucial feature that allows twill robots to be tall
compared to the wheel base dimension. This feature also allows
twill robots to be narrow compared to their wheel base. This
feature results in lack of sensitivity in roll angle of the
surface.
[0059] Other Transitions
[0060] Further work will reveal many other choices for this
transition. We know by example, that a rider on a "fixie bicycle"
sets the steering-angle at about 45 degrees, a slope of just 1.
This may be optimal, convenient, or just an artifact of the fact
that the front-wheel is not powered on a bicycle.
[0061] Front Vs Back
[0062] Please note that the use of "front" and "rear" is purely for
discussion clarity. In fact, twill robots may or not be symmetric
front to rear. In a preferred embodiment the front and rear
steering and traction motor assemblies are identical. This reduces
part count and simplifies other considerations such as the control
laws. For example, if only one-wheel steers, then the twill bot
cannot enter into a true Segway.TM. mode. With one-wheel turned and
the other fixed the bot is in a "half Segway.TM." mode. With minor
adjustments to the algorithm, this balances much like a
Segway.TM..
[0063] Wide Wheels
[0064] Generally, we think of a wheel as touching in a single
point. In practice, a wheel may touch in a considerable round or
oval patch. These offer a small bit of stability while stopped on a
hard surface with little to no roll. The same can be said for a
pair of wheels that are rather close to each other. Even without
articulation, such essentially wide wheels may be used with the
balancing invention presented herein.
[0065] Turning
[0066] Similar algorithms will be required for steering through a
commanded path. The variety of solutions is rich. In a preferred
embodiment, the front-wheel only is steered, much like a bicycle.
In other preferred embodiments, both wheels steer together or in
proportion to each other. It is also possible to drive in
Segway.TM. mode. Remarkably, it is possible to drive in any
direction oblique to the stiff axis and to continuously vary this
axis.
[0067] Hockey-Stop
[0068] In a preferred embodiment, when a transition is required
from forward motion to stopped, one or both of the steering-angles
may be abruptly commanded to 90 degrees. This may result in a
sideways slide of the wheel or wheels much like the common
"hockey-stop" known to ice skaters. Indeed, this may be a universal
transition method if one specifies a minimum speed at which to
command the stop. While sliding and when stopped, any resulting
roll error is handled as a disturbance input.
[0069] This must be kept within the dynamic capabilities of the
servo loops. An open-loop pre bump may be employed to help mitigate
the disturbance of the transition.
[0070] Other Motor Configurations
[0071] It will be clear to one skilled in the art that various
combinations of one-wheel steered or two-wheel steered or one-wheel
powered or two-wheels powered. Indeed, there are other applications
with more than two-wheels in line forming an articulated chain.
[0072] Steering, Encoders
[0073] A natural mechanism for the steering is a fork and headset
as is common on bicycles. This is readily driven by an arrangement
of direct-drive motors, gears, belts, or pulleys as are common in
the art. Steering-angle is readily encoded by any of the
technologies in common use such as magnetic, optical, or others
known to one skilled in the art.
[0074] Non-Zero Trail, Non-Zero Steering-Angle
[0075] A feature of the bicycle steering mechanism is the freedom
to select the steering-angle and trail distance. This geometry has
an effect on the handling and stability of a bicycle and Twill
robots. Future investigation will yield a variety of geometries
suitable for each application. In the context of bicycles this is
still an active area of academic investigation. In Twill robots,
certain geometry selections will reduce control energy and
smoothness of motion.
[0076] Steering Trail for Balancing
[0077] In a preferred embodiment substantial trail distance on one
or both of the steering axes may be used to effect balance control
while stopped without using the traction motors. An extreme version
of this is to have a substantially central hinge that serves as the
steering axis.
[0078] Four-Bar Steering
[0079] An interesting steering geometry was used on some early
motorcycles. This method employs a four-bar mechanism. The key
motivation for this mechanism is to keep the mechanism compact and
generally lower than the top of the wheel. A schematic view of the
mechanism as historically practiced is shown in FIG. 3. Note that
the linkages (22R, 22L) generally connect on the axel (20) of the
wheel (21). At the rear end the linkages connect to the body (23)
with rotational joints. Rod end bearings (24) are commonly used for
these joints.
[0080] Zero Positions for Steering, Roll Etcetera.
[0081] In developing the four-bar steering mechanism for Twill
robots, it has become clear that there is a richness to the
mechanism that has not yet been exploited. To one experienced in
the art of four bar linkages, it is natural to exploit the instant
center of the mechanism. As shown in FIG. 3 this point (31) is at
the intersection formed by the lines defined by the moving links
(22). This is shown schematically in the top view FIG. 3. This
point has the unique property that if the contact point of the
wheel (30) is located there, then steering input results in a pure
rotation about this point. This is only true for small motions, but
is often adequate for practical purposes. In FIG. 3 the contact
point (30) is shown considerably behind the instant center
(31).
[0082] Caster in Four Bar Linkage
[0083] While the above is a natural choice and suitable for some
preferred embodiments, it has an unfortunate property of being
unstable in forward motion. From FIG. 4 we see that for any
non-zero steering-angle, the forward motion force on the wheel
produces a torque on the steering mechanism which acts in a
direction to increase the steering-angle. This could lead to
inherent steering instability. This is equivalent to a negative
caster distance on a castered wheel. This could be stabilized by a
suitably stiff control system. This instability is diminished, but
remains if the contact point is placed below the midpoint of the
moving linkage (as is seen in historic examples). In practice,
moving back from the axel as much as the instant center
substantially eliminates the negative caster.
[0084] Caster Effects
[0085] In each of the cases above, the control law needs to take
into account the effect of moving the wheel off-axis. That is, with
any non-zero caster, moving the contact point of the wheel off-axis
changes the zero-roll angle position. This is readily accounted for
in the control software.
[0086] Suspension, Compact
[0087] The four-bar arrangement results in a compact potentially
self-contained package. In a preferred embodiment this module
contains the drive motor, steering motor and drive electronics.
This greatly enhances field serviceability. In a preferred
embodiment, the package is mounted in a mount that allows
translation or rotation. Adding a spring damper externally to the
module creates a suitable suspension system.
[0088] Hydraulic Synchronization
[0089] In another preferred embodiment, the dual arms of the
four-bar linkage (50, 51) suggest a location for suspension
elements (52) as suggested by FIG. 5. Unfortunately, this
embodiment still requires a means for synchronizing the motion of
both arms. A novel way to accomplish this is to use hydraulic
synchronization as shown in FIG. 6. This is achieved by cross
connecting a pair of double-acting hydraulic cylinders (52L, 52R)
with hydraulic hoses (60) and (61). So, if one cylinder moves, the
other must move precisely the same distance thus synchronizing the
two sides of the linkages. As is common with hydraulic cylinders,
the cross-sectional area on each side differs by the area of the
piston shaft (62). Thus, motion of the piston results in a change
in volume. This is readily accommodated by a pair of auxiliary
cylinders (63) that capture the difference. Coupling these together
with plate (64) preserves the synchronization. In a preferred
embodiment, preloading the pair of auxiliary cylinders with an
external spring (65) provides the suspension spring function. A
flow restrictor (66) on the hydraulic lines may be selected to give
the desired damping forming a high-performance suspension system.
Note that all of this is readily contained in the module further
enhancing field serviceability.
[0090] Knee Jerk Steering
[0091] The steering fork, (25) in FIG. 2 is used in the four-bar
steering to point the wheel in the correct direction. This may be
done by any of the conventional means known to one skilled in the
art. These include but are not limited to electric motor with a
crank or a belt drive.
[0092] Steering Fork Motion
[0093] This is a bit tricky if the suspension allows the fork to go
up and down. In one preferred embodiment a pair of hydraulic
cylinders are used. Through hydraulic connections similar to the
hydraulic synchronizer, the commanded relative position may be kept
constant. Thus, as the fork goes up and down, the commanded angle
can be kept substantially constant.
[0094] Crossed Arms Eliminates Fork Motion
[0095] In a novel four-bar application, the vertical parallelogram
motion is replaced by a four-bar linkage. Generally, the spacing of
the ends (70, 71) may be selected for any desired behavior. The
instant center of the wheel assembly is again where the two lines
cross. So, in this case a preferred embodiment is to cross the arms
as shown in FIG. 7. With the linkage dimensions (72,73) selected so
that the cross is near the attachment point (74) of the steering
fork, the vertical motion has minimal impact on the attachment
point. This substantially reduces bump steering.
[0096] Hydraulic Steering
[0097] Hydraulically assisted steering is common in automobiles.
Hydraulic steering is common in construction equipment and boats.
These would both seem excessively complicated for a twill robot. A
novel implementation has remarkable simplicity. First note that
with one or two hydraulic actuators for steering, the steering
position is commanded by pushing fluid into the correct chamber
while draining the other. This is normally done from a pressurized
hydraulic source. Instead consider moving the fluid directly with a
small gear pump that can move fluid directly in one direction or
the other. A very small gear pump, and large diameter cylinders is
exactly equivalent to selecting a gear ratio in a more conventional
steering train. A careful selection results in a gear free
design.
[0098] In many systems this would seem crazy, since the gear pump
will have a bit of blow-by causing position drift. But in the Twill
robot, the steering position is continually monitored and the
position updated with a servo loop.
[0099] Kneeling and Kick Stands
[0100] All of the foregoing has been accomplished without the use
of kickstands or training wheels. There are preferred embodiments
where kickstands or the equivalent may be included. In one
preferred embodiment, the suspension system spring is an air
chamber compressing the hydraulic differential. Relieving the
pressure allows the suspension to relax-dropping the robot to the
ground. This is similar to the system used on Citroen cars in the
1970's. This is of particular utility in delivery robots that
effectively "kneel" to be loaded or unloaded.
[0101] Tall Thin Delivery Bot
[0102] All of this comes together to form a preferred embodiment of
a delivery robot. As with the other Twill robots it has front and
rear in-line wheels (25, 85). The tall stance (86) allows them to
be as tall as cars on the road. They may have a narrow dimension
(82), to allow lane splitting and driving on the side of the road.
In in FIG. 8 the profile (80) is shown as a frustrated triangle,
but any shape may be used. This thin embodiment is particularly
well suited for delivery of boxes. At this time, a sensible
internal width is only 16 inches as the largest standard Amazon box
is only 16 inches wide. This makes the delivery bot suitable for
driving on sidewalks, through doors, and in freight elevators.
[0103] It is also noted that in a preferred embodiment large
diameter wheels are used. This allows traverse of curbs and even
stairs. As a practical matter, the choice of large diameter wheels
reduces demands on the wheel bearings, provides for long lived
tires. This simplicity results in a vehicle capable of a long
service life with minimal maintenance.
[0104] Arms
[0105] For applications where self-loading and unloading are
required, a variety of embodiments will be fitted with integral
arms (90) as seen in FIG. 9. In one preferred embodiment, the first
link of an arm (91) is formed by a wall of the enclosure. A second
link (92) hinged substantially at the end of the first allows
reaching into the hold to select the next package (95) with a
gripper (93). This configuration allows reaching up to a second
story landing for home delivery. Sensors (94), in particular
cameras, may be included on the arm to insure and document proper
operation and successful delivery of each package.
[0106] Conveyor
[0107] In other preferred embodiments, an internal conveyor belt is
used to dispense packages in a last in first out (LIFO)
sequence.
[0108] Doors
[0109] In other preferred embodiments a door or doors may secure
internal delivery volumes. Upon arrival to the destination, the
appropriate door may be opened remotely presenting the package to
the customer. As is common, a user interface or cell phone
interface may be provided for communication between the customer
and the robot and or between the customer and a remote
operator.
[0110] Autonomy Hardware
[0111] FIG. 8 also shows a LIDAR unit (81) on the top which is
currently common in experimental self-driving cars. The field is
exploding with hardware and software for automatic operation. Twill
robots are suitable for all of these technologies. Sensor and
computation hardware will also benefit from a modular approach to
simplify upgrades and field servicing (96).
[0112] Shorter for Mule Application
[0113] While much has been made of the ability of making twill
robots tall, this does not preclude the ability to make short twill
robots. One application for this configuration is a military style
"mule" (100) for delivering materials as seen in FIG. 10.
[0114] In this application a Twill mule could be used to deliver
materials along trails wherever people walk. The low stance and
narrow width (101, 102) would limit the visibility while traveling.
Polygonal surfaces (103) would future enhance the stealthy nature
of the robot. For manual loading and unloading the top (104) opens
for full access to the hold.
[0115] Telescoping Height
[0116] In one preferred embodiment, a short Twill mule can navigate
to an area of interest. Instead of delivering materials, the
payload could be a telescoping device such as but not limited to a
camera or antenna (111) as shown in FIG. 11. Taking full advantage
of the twill robot agility, a tall surveillance mast (110) may be
fully deployed while driving providing a clear view over walls or
over a tree line.
[0117] Telepresence Robot
[0118] A natural application of tall skinny Twill robots is for
telepresence robots. One preferred embodiment is shown in FIG. 12.
This is a growing industry that uses conventional platforms and
Segway.TM. style platforms. Twill telepresence robots can telescope
up and down while in use to adjust conversation height (120).
Completely collapsing the legs (121) results in a compact package
suitable for convenient shipping to a remote location. The thin
package dimension (122) is ideal for sharing space with humans in a
hallway. A multiplicity of cameras (124) serve both for navigation
and sending video to the remote station. To have a conversation,
the telepresence Twill turns to face the audience presenting a full
display (125). The other side may have a display also.
[0119] Surveillance Robot
[0120] Another application for tall and skinny twill robots is for
automated surveillance robots. This is a market that has been
limited by plodding quasi static robotic platforms. Twill
surveillance robots mix better with foot traffic, have a
vanishingly small frontal profile (to avoid collisions and
detection).
[0121] In Closing
[0122] It will be clear tone skilled in the art that there are many
additional applications for Twill robots.
[0123] In view of the above teaching, a person skilled in the art
will recognize that the apparatus and method of invention can be
embodied in many different ways in addition to those described
without departing from the principles of the invention. Therefore,
the scope of the invention should be judged in view of the appended
claims and their legal equivalents.
* * * * *